A Novel Sensor Concept for Selective and Self-Powered Gas Detection

A Novel Sensor Concept for Selective and

Self-Powered Gas Detection

Martin W. G. Hoffmann  

Aquesta tesi doctoral està subjecta a la llicència Reconeixement 3.0. Espanya de Creative

Commons.

Esta tesis doctoral está sujeta a la licencia Reconocimiento 3.0. España de Creative

Commons.

D

T

P

D

D

U

B

A Novel Sensor Concept for Selective and

Self-Powered Gas Detection

Martin W. G. Hoffmann

T

D

Dr. Juan Daniel Prades García

Dr. Francisco Hernández-Ramírez

T

Prof. Albert Cornet i Calveras

F

ACULTY OF

P

HYSICS

D

EPARTMENT OF

E

LECTRONICS

MIND

G

ROUP

,

IN

UB

P

D

E

T

A

A Novel Sensor Concept for Selective and

Self-Powered Gas Detection

Tesi que presenta

Martin W. G. Hoffmann

per optar al títol de Doctor per la Universitat de Barcelona

D

T

Dr. Juan Daniel Prades García

Dr. Francisco Hernández-Ramírez

T

Prof. Albert Cornet i Calveras

F

F

D

´E

G

M

-N

D

E

F

(MIND)

i

Resum

Les tecnologies de sensors de gas basades en semiconductors presenten limitacions importants de selectivitat i consum d'energia. Per tant, esdevé necessari l’assaig de nous conceptes de dispositius capaços de satisfer aquests dos requeriments per aplicar-los en plataformes mòbils. En aquesta tesi es presenta una tecnologia de sensors altament selectiva i autònoms des d’un punt energètic, incloent la seva avaluació experimental i l’anàlisi dels mecanismes físico-químics de detecció subjacents. S´han fabricat materials nano-híbrids, basats en nanofils inorgànics (NWs), funcionalitzats amb monocapes “auto-acoblades” (SAMs). A la dissertació es mostren les extraordinàries característiques en termes de selectivitat i sensibilitat de gas que exhibeixen aquests materials; els estudis teòrics són consistents amb les observacions experimentals disponibles i permeten identificar l'estructura electrònica dels orbitals moleculars de frontera SAM, que és el paràmetre crucial per a garantir una interacció eficaç entre el sensor i els gasos. A més a més, es presenta un nou concepte de sensor autònom sobre la base d'una heteroestructura p-Si/n-ZnONW que respon exclusivament a la llum solar sense necessitar d’altres fonts d'energia externes. Els canvis de la tensió de circuit obert (Voc), que s’utilitzen per controlar la

presència d'espècies de gasos, mostren una correlació directa amb la densitat de portadors de càrrega (Nd) al nanofil de n-ZnO. Finalment, es presenta l’aplicació de tècniques de

microfabricació en el disseny d’un dispositiu que integra els conceptes de selectivitat i autonomia energètica, capaç per tant de detectar concentracions de NO2 rellevants per a

aplicacions de seguretat (nivell de ppb) sense la necessitat de fonts d'energia externes. La mida compacta, la baixa demanda d'energia i la robustesa de la tecnologia fan que el concepte de sensor que aquí es presenta sigui molt prometedor per a la seva integració futura en plataformes electròniques mòbils.

Abstract

Contemporary semiconductor based gas sensor technologies could already prove their high sensitive characteristics but exhibit crucial debilities in terms of target selectivity and power consumption. As both criteria have to be fulfilled for the application in mobile sensor platforms, new device concepts are needed. Within the here presented thesis the development of a highly selective and self-powered as well as experimental evaluations and analysis of the underlying sensing mechanisms are presented. First, hybrid nano materials are fabricated, based on inorganic nanowires (NWs) functionalized with self-assembled monolayers (SAMs), and show extraordinary characteristics in terms of gas selectivity and sensitivity. Theoretical mechanistic studies are consistent with the experimental observations and identify the electronic structure of the SAM frontier molecular orbitals as crucial parameter for efficient sensor-gas interactions. Furthermore, a novel self-powered sensor concept is presented based on a p-Si/n-ZnO NW heterostructure that is solely driven by solar light and without the need of external energy sources. Changes of the open cirquit voltage (Voc) that are used to monitor the presence of gas species are shown to correlate

with the charge carrier density (Nd) within the n-ZnO NW upon gas-sensor interactions.

Finally, microfabrication techniques are applied to unify the selective and self-powered concepts within a single sensor device that is capable to selectively and quantitatively detect NO2-gas concentrations within security relevant concentrations (ppb level) and without the

need of external energy sources. The compact size, low energy demand and validity of signal information make the here presented sensor concept very promising for the integration into mobile electronic platforms.

Acknowledgement

The realization of the here presented thesis goes along with the handling of various scientific and technical disciplines that could never be solved alone by myself in their entirely. Therefore, this acknowledgement is dedicated to all persons that helped and supported me within the years in Cologne, Barcelona and Braunschweig.

At first, I want to give very special thanks to my thesis supervisors Daniel Prades and Francisco Hernandez-Ramirez, as well as Hao Shen for their extraordinary support and inspiration during all my thesis work in scientific fields and far beyond. Thank you for showing me the joy and confidence to work on new and exciting projects that allowed me to learn a lot about their physical background. It was a pleasure to work with you during this thesis! Thank you very much Dani for your great hosting during all my visits and for showing me a part of the real Barceolona.

Also, I want to give a special thank to Andreas Waag for giving me the chance and trust to take part in the development of exciting strategic projects. This fruitful and inspiring collaborations always had a high value for my personal development.

Alaaeldin, thank you very much for being my friend and for sharing all ups and downs during this time. I really appreciate to work with you and to develop our -sometimes crazy- ideas. Here, I also want to thank Lorenzo, Dennis and Hao for being my friends here in the far north, and especially Lorenzo for giving me the chance to lose various kicker-table bets. I am really thankful to share thoughts with all of you inside-, and especially outside the lab. This thesis would not have the same scientific value without the excellent theoretical models developed by Leonhard Mayrhofer from the Fraunhofer IWM in Freiburg. Thank you very much for this inspiring cooperation that enabled us to get an idea about the complex mechanism of selective gas detection. I also want to thank Tommi Järvi and Michael Moseler for the support of these works.

As mentioned above, this thesis could not be done without the support of my colleagues. I want to thank all IHT´ler for their helping hands in many aspects. A special thank to Angelika S. and Juliane for their excellent technical support to realize self-powered sensor devices. Sincere thanks to Angelika J. for the great organization of various urgent needs and Klaudia for the support in the LENA project. Thanks to Andrej, Alex, Frederik, Sönke and Ito for world class football matches, Jana and Xue for providing high quality GaN strucutures, as well as all the colleagues in Braunschweig: Andrey, Doris, Erwin, Feng, Hergo, Jandong, Johannes, Mahmud and Xiaodan.

v I want to thank the fantastic colleagues from the MIND group: Olga for giving me a deeper insight into semiconductor physics and the quality of red wine, Jordi for showing me the easy handling of E-beam lithography facilities, as well as Oriol and Giovanni for many helping hands in the lab. Thanks for feeling home in Barcelona.

Alexandra, Daniela, Jan, Kai, Micha, Rainer, Robin, Stefan and Thomas; thanks a lot for the long journey and your friendship. You are great!

Am Ende möchte ich den für mich wichtigsten Menschen danken: Britta für die einzigartige Freude, Motivation, Unterstützung und Geduld während der gesamten Arbeit sowie Mama, Jan, Simon Rike, Justin und Bärbel. Ich hoffe, dass ich Euch in dieser Zeit trotz aller Gedanken um diese Arbeit Eure Bedeutung für mich zeigen konnte.

Content

1 INTRODUCTION

FEHLER! TEXTMARKE NICHT DEFINIERT.

1.1SELECTIVE SENSING 1

1.2SELF-POWERED GAS SENSING 9

1.3DISSERTATION OUTLINE 12

2 OBJECTIVES

14

3 RESULTS AND DISCUSSION

16

3.1SELECTIVE SENSING 17

3.1.1 PAPER 1 20

3.1.2 PAPER 2 48

3.2SELF-POWERED SENSING 56

3.2.1 PAPER 3 58

3.2.2 PAPER 4 76

3.3COMBINATION OF SELECTIVE AND SELF-POWERED SENSING 82

3.3.1 PAPER 5 84

4 CONCLUSION

116

5 OUTLOOK

120

6 REFERENCES

124

APPENDIX

130

1

Introduction

The past decades were determined by a fast growing earth population and an ambitious industrial growth of emerging economies in Asia and Latin America. A consequent transformation from agriculturally dominated structures to urban societies let metropolises grow to megacities as Mumbai, Shanghai, Istanbul or São Paulo.1 _{So far, a fast economic}

growth was rated with higher priority than the caused environmental impact. A more and more serious pollution of air, water and agricultural areas in urban regions with consequences on human health and even social unrest led to a change of thinking and a political intent for vigorous efforts to target environmental problems within the next decades. As Industrial production sites are mainly concentrated in urban regions, the close proximity to the living environment creates a complex challenge to ensure sustained economic growth in uprising metropolitan areas.2

The development of new gas monitoring systems will play an essential role in programs to realize a higher life quality. A sufficient control of densely populated areas requires close meshed sensor networks with high detection accuracy in terms of location and gas selectivity to identify threads and to take efficient actions. Moreover, people will be interested to identify harmful threads also in their personal environment without the need of a public infrastructure. Mobile devices like smart phones could serve as platforms to integrate the needed sensor technologies and ensure an on-demand accessibility. A low impact on the battery lifetime of such devices and an easy interpretation of the detection results, even for non-specialist users, is essential for a successful commercial application. Therefore, to facilitate wide spread and personal analyses of the air quality, gas sensors

have to accomplish two major requirements:

 high target selectivity

 low power consumption

Both qualifications cannot be sufficiently fulfilled by current sensor technologies.3

Electrochemical or optical analysis systems can offer a high specificity, but their bulky

2-1 1 Introduction dimensions, as well as high power consumption and costs do not allow them to be applied for mobile or wireless applications.4,5 _{Resistive gas sensors, based on semiconducting metal}

oxide materials, with their relative small size, simple setup, high sensitivity, robustness and low costs have shown a high potential on the way to be implemented into common silicon based platforms. For metal oxide materials, the gas interaction process proceeds via a gas adsorption at lattice oxygen sites (Olat) or terminal hydroxyl groups (-OH) on the

semiconductor surface.6,7 _{These sites can be described as target receptors.}8 _{Depending on}

the oxidative or reductive character of the gas species, the charge carrier density (electrons or holes in case of n- or p-type materials respectively) within the material surface area is modulated via a chemical interaction with intrinsic oxygen vacancies (V_O) as described for oxygen by equation (1):

(1)

Nano structuring of sensing materials led to significant improvements in terms of sensitivity due to high surface to volume ratios and defined conduction channels.9,10 _{However, the}

unspecific adsorption and redox interaction of the native oxygen receptors cannot discriminate specific gases. Additionally, high activation energies, delivered by external heaters or UV light sources, are needed to facilitate the gas-surface interactions and create a measurable signal to monitor the surface processes (e.g. voltage or conductance).

So far there are no specific concepts to overcome these limitations. Therefore, this thesis is devoted (a) to develop a new pathway for gas selective semiconductor based gas sensors beyond current technologies and (b) to realize a sensor concept that is capable to operate without the need of external energy sources. Finally, both principles should be implemented in one device to accomplish both criteria of selectivity and low power consumption in one material system.

1.1 Selective Sensing

Previous works towards the development of gas-specific “electric noses” were concentrating on inorganic modifications of metal oxides with noble metals,8,11 _{chemically reversible}

hetero materials, as CuO 12 _{or Ag2O} 8 _{or statistical analyses of sensor arrays.}13–15

Nevertheless, these devices go along with restricted diversity of targeted gas species or a high system complexity. In contrast to those systems, the human olfactory system is based

2-1.1 Selective Sensing 2 on organic structures and can discriminate up to one trillion distinct odors, even at concentrations in the range of parts per billion (ppb).16

Fig. 1.2: The organization of the human olfactory system with schematic illustration of odorant

receptor proteins as molecular binding sites (reprinted from The Nobel Assembly at Karolinska

Institutet).17

The recognition of certain molecular specie begins with a binding interaction of a receptor protein and a consequent signal transduction trough the receptor cells to the brain. A mimicking concept with receptor proteins or antibodies and inorganic nanowires (NWs) as signal transduction paths has been successfully applied for the detection of biological targets in the liquid phase.18–20 _{Even selective orthogonal on/off response characteristics}

could be achieved. Recent works indicate that organic surface modifications could also play an important role as specific and flexible receptors for selective target discrimination in the gaseous phase.21–24 _{In contrast to these works introduced until today, the obtained studies of}

3 1 Introduction the here presented thesis indicate that not only the selective binding interaction between the organic receptor and a gas species, but more importantly the mechanism of the charge transfer process between the organic surface molecules and the inorganic semiconductor material are critical for a selective sensing. Therefore, a brief insight into these organic-inorganic interface charge transfers will be given prior to a discussion of the relatively new field of hybrid gas sensor systems.

Electrical modulation of semiconductors by organic surface molecules

An obvious method for the deposition of organic components on inorganic materials is the plain physical vapour deposition (PVD) under vacuum conditions.25,26 _{The control of the}

layer thickness and conformal deposition, however, is not easy to control and requires high analytical efforts and optimization. Alternatively, self-assembled monolayers (SAMs) represent a very convenient way to modify the surface of semiconducting materials with organic functionalities. SAMs are formed by the adsorption of molecular constituents from the gas phase or solution. These molecules possess a terminal chemical functionality with high affinity to bind with the material surface (head group) and a tail with desired chemical composition (see figure 1.3).26,27 _{Binding to material surfaces can occur e.g. via condensation}

reactions (as in case of reactions of silanes with metal oxides), hydrogen substitution (Thioles with gold) or π-π interactions (terminal aromatic groups with carbon nanotubes). The saturation of the material surface after adsorption of the first organic monolayer prevents the biding of further monomers, as no more free sites of the semiconductor surface are accessible.

Fig. 1.3: Schematic description of SAM formation on semiconductor surfaces and exemplary

redpresentation of surface affine head groups for immobilizations on metal oxides, gold,26 _and CNTs.28

When organic molecules get in contact with a surface of inorganic materials, the electrical properties of the semiconducting material can be considerably changed. The use of nanomaterials, (e.g. quasi 1D NWs) is beneficial for this interplay, as the reduced radius of

1.1 Selective Sensing 4 these structures causes a higher fraction of the surface area (S) in comparison to the volume of the bulk material (V):

r l r rl V S 2 2 2  





Here, r is the radius and l the length of the NW. It is evident from equation 2 that the surface to volume ratio (S/V) is increased with decreasing NW radius. Consequently, the SAM modified surface has a stronger influence on the overall material properties for nano sized materials compared to their macroscopic counterparts.

Beside the size effect of the NW, the donor/acceptor properties of the organic counterpart, as well as the bulk work function are responsible for the materials electrical modulation.29

Kong et al. already reported in 2001 that the adsorption of amine containing organic SAM, formed by the surface immobilization of 3′-(aminopropyl)tri-ethoxysilane (APTES), had a considerable influence on the electrical conductance of semiconducting p-type carbon nanotubes (CNT).30 _{It was proposed that these changes were induced by electron donation}

from APTES to p-CNTs and a consequent reduction of the p-type charge carriers. This effect was described as chemical gating. Although a lack in theoretical understanding of the underlying charge transfer mechanism was mentioned in this study, this term is still used to describe the electrical response of organic surface modified biosensors after target binding.19,31

Qi et al. could develop a model to describe the electronic states and charge transfer interactions of strong organic electron acceptor groups and intrinsically undoped bulk materials before and after a direct contact or binding interactions.25,32 _{The consequent}

phenomenon of induced bulk conductivity is known as surface transfer doping. In this study, intrinsically non-conductive hydrogenated diamond (diamond:H) and the strong organic electron acceptor tetrafluoro-tetracyanoquinodimethane (F4-TCNQ, el. affinity: χ = 5.24 eV)

were used. As F4-TCQN has no affine head group, the molecules were deposited by simple

evaporation. Photoemission spectroscopy (PES) experiments of core levels (CL) and valence regions (VR) before and during the organic deposition clearly demonstrated shifts of the intrinsic energy levels that were interpreted as upward bend bending at the diamond:H surface and the formation of an interface dipole with negative charge on the electron accepting organic group. In other words, these findings described a charge transfer from diamond to the electron affine F4-TCNQ, resulting in a p-doping close to the interface, as an

5 1 Introduction

Fig. 1.4: Schematic energy diagram of bulk material (hydrogenated diamond) and an organic electron

acceptor (F4-TCNQ) (a) before and (b) after surface transfer doping.32 _{(c) Chemical structure of} F4-TCNQ.

A critical requirement for an efficient charge transfer from bulk to the electron acceptor is a lower energetic position of the lowest unoccupied molecular orbital (LUMO) of the organic acceptor with respect to the bulk Fermi energy (EF). The electron transfer proceeds until the

LUMO energy position aligns with the bulk EF. Notably, no further charge transfer was

observed with ongoing acceptor deposition after the first organic monolayer.

The same effect of surface doping by an organic surface functionalization was observed for bulk organic semiconductors, as single crystalline rubrene, when fluorinated SAMs as (tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane (FTS) were attached to their surface.33

A conductivity increase as big as five orders of magnitude could be observed in such a system with respect to the unmodified material.

Fig. 1.5: a) In situ semi-log plot of the source-drain current for organic semiconductors (rubrene,

tetracene) during SAM deposition. Fluorinated (FTS) and non fluorinated SAMs (Ethoxysilane, OTS) were deposited at t = 0. b) Electron spin resonance (ESR) spectra of pure and combined components (semiconductor, SAM). The pronounced resonance indicates unpaired electrons as well as delocalized holes for the rubrene/FTS-SAM system. (Reprinted by permission from Macmillan Publishers Ltd: Nat. Mater [33_{], copyright 2008.)}

a

b

c

1.1 Selective Sensing 6 As in the case of diamond:H, the strong electron withdrawing ability of the SAMs lead to an efficient charge transfer, leaving unpaired electrons (transferred to the SAM), as well as delocalized holes (charge carriers in the semiconductor) in the semiconductor (see figure 1.5).

From the above mentioned studies it can be concluded that organic monolayers can strongly influence the electrical properties of semiconducting materials by modulating their charge carrier concentrations close to the bulk-organic interface. For this reason, the use of nano structured semiconductors with high surface to volume ratio is beneficial to achieve a strong deviation of the initial electronic properties. The electron affinity (χ) of the organic monolayer, which is defined by the LUMO position with respect to the vacuum level (Evac), is

the most critical parameter for an efficient charge transfer. It is shown in the here presented thesis that binding interactions of gas molecules to SAMs modulate the LUMO energy of the organic surface molecules. A mechanism how these modulations can lead to a selective response of a SAM-NW sensor system is described in the papers 1 and 4.

Recent developments in the field of SAM-NW hybrid bio/gas sensors

Organic surface modifications were used first to gain selectivity for NW based biosensors in liquid phase. Antibodies,19 _peptides,34 _{as well as single stranded DNA}35 _{were used as}

receptor groups. However, the detailed charge transfer mechanism between the organic surface component and the semiconductor antenna was not taken into account systematically, but the selectivity was meant to be exclusively induced by selective target-receptor binding and consequent changes of the NW surface potential (Ψ0). The relationship

between Ψ0 and the charge (σ0) of the sensor conduction channel (here the NW) is usually

described by the Graham equation:3

        kT e kTCo o W o 2 sinh 8







Where k is the Bolzmann constant, T is the temperature, e the elementary charge, ε0 the permittivity of free space εw the dielectric constant of water, and C0 is the ionic strength of

the buffer solution. In the case of bio sensing, a wide know how about these key-lock interactions from well understood fields of biotechnology could be transferred to new nanotechnology applications. Antibodies can be designed as counterpart for defined antigens, as cancer marker proteins or viruses, by methods as recombinant in vitro antibody production techniques.36 _{As the chemical and sterical structure of a targeted antigen}

possesses a number of specific binding sites and characteristic structural properties, which are reflected in the structure of the adequate antibody, a very specific binding interaction can be achieved.37–39 _{The first NW biosensor based on this concept was described by Lieber}

7 1 Introduction

et al. in 2001 for the biotin/m-antibiotin target/receptor couple and silicon NWs as electric

channel material.40 _{Furthermore, the same group could develop this system to realize}

antibody functionalized NW sensors that showed an extraordinary selectivity towards respective parostate cancer markers in concentrations as low as 0.9 pg/mL. The fact that these measurements were performed using undiluted serum samples indicated the high potential for diagnosis applications.19 _{The principle of key-lock interactions is also used for}

the detection of single stranded DNA (ssDNA) by immobilizing DNA strands with a respective amino acid sequence to achieve specific DNA hybridization.41,42

Fig. 1.5: Schematic illustration of biosensor concepts based on specific a) antibody-antigen binding

and b) hybridization of immobilized receptor- and target DNA strands.

These hybridization interactions can be monitored with a very high accuracy and low noise level. Therefore, it was possible to perform even single molecule measurements with high time resolutions in the micro second range,, using a point defect CNT based FET system. Beside specific on/off detection signals, even characteristics of the molecular interaction kinetics, as temperature dependent dwell-time of target ssDNA, could be studied.35,43

The use of organic receptor molecules for biosensors is evident, as these immobilized species possess similar sizes, functionalities and atomic structures compared to the targeted antigens. On the same time, organic structures can be realized in an infinite number of variations of primary or secondary structures to accomplish the desired functionality. Learning from natural olfactory systems, the use of defined organic receptors, also for small molecules in the gas phase, seems to be a very promising route to solve current selectivity issues of pure inorganic semiconductor sensors. Up to date, just a few works have been presented to study the potential of well defined organic/inorganic hybrid materials for gas sensing applications. In 2008, McAlpine et al. used silicon nanowires decorated with known recognition peptides for the detection of acetic acid and ammonia vapours in dry nitrogen.44

For relatively high concentration of 100 ppm, a remarkable discrimination could be observed in presence of interfering gases, as CO2 and acetone. As the peptide-target

interaction causes a protonation or deprotonation of the peptide in case of acidic acid or ammonia, respectively, the sensor signal was found to be not reversible. Additionally, the

b

a

1.1 Selective Sensing 8 absence of stabilizing buffer solution, as in case of biosensing experiments, led to insufficient peptide stability in presence of even very low humidity values. Possibly, the loss of the protein secondary structure after water adsorption caused the loss of the specific binding ability with the respective target molecule.

Fig. 1.6: Different approaches for the realization of gas sensors with organic receptor units:

a) Covalent immobilization of tailored ammonia or acetic acid binding peptides;44 _{b) metal-organic} Cu(I) complexes as receptors for none polar ethylene molecules23 _{and c) pure organic thiourea SAMs} for the detection of cyclophenon vapours.45

A direct transfer of hybrid biosensor concepts to gas sensors by using tailored protein modifications seems to have stability limitations, as the target-receptor interactions take place in none physiological conditions. On the other hand, the small dimension and lower structural complexity of gas molecules, compared to biological antigens, lowers the influence of shape controlled key-lock interactions with receptors. Very recently, pure chemical interactions with receptors in sizes comparable to gaseous targets and secondary structures were found to be very powerful receptors and more stable in gaseous environments. Esser et al. used single walled CNTs (SWCNT) modified with organo-metallic Cu(I) complexes for the detection of small and none polar ethylene molecules in the sub-ppm range and with a noticeable selectivity.23 _{Other than SAM systems, the metal-organic}

complex molecules were not covalently bond to the CNT surface, but just mixed in solution and the mixture was drop-casted on an interdigitated electrode. Possibly, the loose interaction between CNT and Cu(II) species led to relatively low sensitivities (sensitivity ~1% for 50 ppm). This system mimics the natural ETR1 ethylene binding receptor, but concentrates only on the essential Cu(I) binding site of the complex protein structure. Such strategies are well known and established in the development of metal-organic and pure organic catalysts,46–50 _{as further reduction of the receptor construction leads to purely}

organic functionalities. Such systems have been demonstrated for the detection of volatile

9 1 Introduction organic compounds15,51,52 _{and explosives.}45,53 _{Nevertheless, although remarkably higher}

system sensitivities compared to unmodified materials could be achieved, these systems could not deliver satisfying selectivity towards singular gas species, unless complex pattern recognition algorithms were applied. Up to date, all surface modified hybrid gas sensors systems have been studied empirically. Especially for pure organic SAMs, the chemical and electrical interactions of the NW-SAM-gas systems are not well understood. Schorr et al. very recently tried to understand the binding interaction between thiourea functionalized CNTs with cyclohexanone targets by nuclear magnetic resonance (NMR) studies (see figure 1.7c).45 _{A direct correlation with interactions in the gas phase is not evident, as these}

experiments were performed in liquid phase and solvent effects can influence the interaction of both components.54 _{The systematic development of tailored hybrid gas}

sensors -selective towards specific gas species- will need systematic studies not only on binding interactions, but also on the electronic structure that can result in efficient charge transfer and resulting changes in the monitored NW resistivity. The here presented thesis, therefore, is aiming to develop highly selective hybrid sensor systems on one hand, and understand the underlying sensing mechanisms (see papers 1 and 4). The reduction of complexity, in terms of the organic SAMs structure, enables the use of theoretical density functional theory (DFT) methods, and gives insights into molecular and electronic interaction mechanisms that are far beyond the resolution of current experimental analytics. This understanding of the NW-SAM-gas charge transfer and gas-SAM chemical interactions, combined with experimental sensing results will be further used to develop conditions for optimized device operation.

1.2 Self-Powered Gas Sensing

The last two decades of technical development where mainly dictated by groundbreaking innovations in the field of information technologies. Today, there is a wide availability of powerful and compact mobile devices for communication, identification, positioning, tracking or management.55 _{Especially within the last few years, the integration of such units}

in the “internet of things” enabled the correlation of multiple parameters from large areas to gain complex information and a new level of informative value.56,57 _{As mentioned earlier, the}

integration of selective gas sensors in such platforms and their connection in wide spread networks would open the possibility to monitor the atmospheric quality for areas of interest with high local resolution and not only at discrete spots. However, the used sensors must be operated self-sustained and maintenance free conditions to ensure an affordable and

1.2 Self-Powered Gas Sensing 10 continuous application. The major requirement is a minimal power consumption of the active device to ensure applicable battery lifetimes and costs per piece. Within the past five years, first concepts have been developed to remarkably reduce the energy demand of sensor elements58 _{or to integrate separate energy harvesting components with sensors to}

accomplish self-powered operation of the combined system.5,59–61 _{Prades et al. made use of}

the small dimension of a single SnO2 NW to induce self-heating at higher applied currents

(>10 nA) and thus, achieved a thermal activation of the semiconductor surface without the need of an external heater (see figure 1.8).58 _{This effect could be attained due to the low NW}

diameter (<100 nm) and a reverse biased SnO2-Pt Shottky contact (deposited by focused

ion beam (FIB) lithography) that posses both a high resistivity (MΩ range). The tiny energy demand (µW range) of the nano sized single NW sensor element further enabled the use of a conventional thermoelectric microgenerator (external energy harvester) to realize a self-powered sensor system.5

Fig. 1.8: (a) SEM image of a single SnO2 NW contacted by FIB lithography and schematic illustration of the device resistance, as well as heat induced by the Joule effect upon currents ≥10 nA. (b) Sensor responses revealing a thermal SnO2 surface activation at applied currents ≥10 nA.58 _{(Reprinted with} permission from [58_{]. Copyright 2008, AIP Publishing LLC)}

The same concept of combining an energy harvesting component and a sensor element in a combined system was used by Wang et al.. Most of these works were concentrating on ZnO nanogenerators as energy harvesting unit.55,60,62,63 _{These nanogenerators use the}

piezoelectric effect, induced by changing the compressive stress applied on ZnO NWs, to harvest mechanical energy from the environment (output power up to P ≈ 31.2 mW/cm³).64– 66 _{Mechanical energies are originated from sonic waves, friction or vibrations and thus are a}

steady energy source to drive sensors in a sustained operation.

11 1 Introduction

Fig. 1.9: Schemes of a) the working principle of a piezoelectric nanogenerator and b) the integration

of sensing and charging units to realize self-powered sensing operations.

By integrating energy storage units, (e.g. Li-ion batteries or super capacitors) as a third component into the sensor system, further stability of the operation could be achieved. These devices can change between a charging and sensing mode. During charging mode, the energy harvesting unit is charging the energy storage unit. For the sensing mode, the storage unit is supplying energy to the sensing unit for detection of the appropriate target.55 _{As the}

self-powered sensor systems are assemblies of individual components, the nature of energy harvesting and sensor units are flexible and can by chosen by the appropriate application. Alternatively to energy harvesting nanogenerators, microbial fuel cells (MFC,

P = 30 µW/cm²)67 _{or combined hybrid cells could be used. The latter are combined systems}

of nanogenerators and solar cells or MFS (Pmax = 100 mW/cm²).61,68,69 These combinations could deliver higher energies to the sensor system, due to simultaneous energy harvesting of different environmental energies. Additionally, the sustained power supply, even in absence of one kind of energy, is ensured and allows for a continuous device operation. Although sensors for the detection of pH-values, UV-light or metal ions could be used for such combined systems, the high energy demand of conventional and low cost gas sensors, that is usually in the order of a few hundreds of mW,70 _{was disabling an applications in}

combined harvester/sensor systems.

To overcome the limitations of self-powered sensors for gas sensing applications, a new concept is presented in this thesis. Different from existing approaches, the developed gas sensor device unifies energy harvesting and sensing unit in a singular multifunctional heterostructure. Thus, there is no need of combining two individual components, with consequent decrease of system costs and complexity. Solar light was used as exclusive energy source to (a) facilitate the gas-sensor interaction and (b) generate the monitored sensor signal. The relevant results are presented in the papers 2 and 3.

1.3 Dissertation Outline 12

1.3 Dissertation Outline

The guideline of the here presented PhD thesis was defined by objectives that are summarized in the following chapter (Chapter 2: Objectives)

The results of the PhD work and their detailed discussion are presented in three sections (Chapter 3: Results and Discussion), namely: 1-Selective Sensing, 2-Self-Powered Sensing, and 3-Combination of Selective and Self-Powered Sensing. Each section is introduced by a short description of the presented content, followed by the main results obtained (as they were published in scientific peer-review journals), and finally the most significant results are summarized with special focus on the classification in the state of the art research.

The consequent chapter (Chapter 4: Conclusion) gives a conclusion of the most significant results obtained during the PhD thesis work

Finally, this thesis will present an outlook on following activities within the topics of selective and self-powered gas sensors. These upcoming research work will be based on the results of the here presented work and should extend the principles to develop selective sensors towards different target gases and realize highly integrated low power sensing platforms.

Only the publications contained in this list shall be considered for the evaluation of this Ph.D Dissertation. A copy of all these publications can be found in the pageindicated.

Section 1:

1. Hoffmann, M. W. G.; Prades, J. D; Mayrhofer, L.; Hernandez-Ramirez, F.: Moseler, M.; Waag, A. and Shen, H. “Highly selective SAM-nanowire hybrid NO2 sensor:

insight into charge transfer dynamics and alignment of frontier molecular orbitals”, Adv. Funct. Mater. 24, 595–602 (2014). (impact factor: 10.4; inside cover 5/2014) Page 23

2. Shao, F.; Hoffmann, M. W. G; Prades, J. D; Zamani, R; Arbiol, J.; Morante, J. R.;

Varechkina, E.; Rumantseva, M.; Gaskov, A.; Giebelhaus, I.; Fischer, T.; Mathur, S.; Hernandez-Ramirez, F., “Heterostructured p-CuO (Nanoparticle)/n-SnO2

(Nanowire) Devices for Selective H2S Detection”, Sens. Actuat. B 181, 130-135

13 1 Introduction Section 2:

3. Hoffmann, M. W. G.; Gad, A.; Prades, J. D.; Hernandez-Ramirez, F.; Fiz, R.; Shen, H.; Mathur, S., “Solar diode sensor: Sensing mechanism and applications“,

Nano Energy 2, 514–522 (2013). (impact factor: 10.2) Page 61

4. Gad, A. E.; Hoffmann, M. W. G.; Hernandez-Ramirez, F.; Prades, J. D.; Shen, H.; Mathur, S., “Coaxial p-Si/n-ZnO nanowire heterostructures for energy and sensing applications”, Mater. Chem. Phys. 135, 618–622 (2012). (impact factor: 2.1) Page 79

Section 3:

5. Hoffmann, M. W. G.; Mayrhofer, L.; Casals, O.; Caccamo, L.; Hernandez-Ramirez, F.; Moseler, M.; Waag, A.; Shen, H.; Prades, J. D., “Highly selective and self-powered gas sensor enabled via organic surface functionalization”,

14

2

Objectives

The objectives of the here presented PhD thesis can be summarized in the following thematic line-up:

1. To develop a gas sensor system that possesses high target selectivity beyond the state of current technologies. This includes the understanding of the underlying mechanism to be able to extend the sensing concept for other target species.

2. To develop a gas sensor concept that can operate in a fully self-powered mode. The system should be designed in a way that it unifies sensing and powering unit in a singular unit.

3. To combine the previous results on concepts for selective and self-powered

gas sensors in one system, to realize a device that fulfils these two major criteria on

16

3

Results and Discussion

In this chapter, the most relevant results of the PhD work and their discussions will be presented in form of papers published in peer reviewed scientific journals. In succession to the above described objectives of the here presented PhD thesis, the papers are organized in three sections:

 Section 1: Selective Sensing

 Section 2: Self-Powered Sensing

 Section 3: Combination of Selective and Self-Powered Sensing

In the first section, a hybrid organic-inorganic sensor system for highly specific NO2

detection is presented. The gas sensing performances achieved by this system, in terms of selectivity and sensitivity, were far beyond the capabilities of NW based systems reported up to date. Further discussion of theoretical DFT simulations of the entire hybrid-gas system enabled the identification of crucial mechanistic parameters to achieve selectivity towards single gas species.

The second section is devoted to the development, mechanistic understanding and application of p-Si/n-ZnO heterostructures. During this PhD work, self-powered gas sensors could be realized using these materials, benefiting from its feasibility to generate a sensing signal and interact with the surrounding atmosphere just by harvesting solar light energy. In the last section of this chapter, the principles of the two previous sections were combined to realize a highly selective and self-powered gas sensing device. Further mechanistic insights for both -selective gas detection by hybrid systems, as well as p-n heterostructures for efficient harvesting- are given in this section. Additionally, a new gas sensor platform was designed and fabricated by micro structuring techniques to ensure high signal intensities and flexibility in terms of inorganic and organic components.

17 3 Results and Discussion

3.1 Selective Sensing

Included paper 1:

Hoffmann, M. W. G.; Prades, J. D; Mayrhofer, L.; Hernandez-Ramirez, F.: Moseler, M.; Waag, A. and Shen, H. “Highly selective SAM-nanowire hybrid NO2 sensor: insight into charge transfer

dynamics and alignment of frontier molecular orbitals”, Adv. Funct. Mater. 24, 595–602 (2014). (impact factor: 10.4; inside cover 5/2014)

Semiconductor based gas sensors are hampering from low selectivities toward specific gas molecules. The reason for this limitation is the underlying sensing mechanism based on the chemical redox reaction between inorganic surface oxygen groups (in case of metal oxide materials) and the surrounding gas atmosphere. In the here presented paper, organic surface functionalities were anchored as SAMs on a SnO2 NW surface to gain highly specific

gas interactions, caused by the organic surface receptors, that can be monitored by the NW resistance. Nitrogen oxide (NO2) was chosen as targeted species, as it presents one of the

major threads for human health, already in low concentrations (ppb level), and is a common pollutant in urban areas.71 _{Different amine terminated methoxysilanes were evaluated as}

SAM receptors towards NO2, as amines (nitrogen oxidation state: –II) were expected to act

as electron donor when electron deficient nitrogen species –as in case of NO2 (nitrogen

oxidation state: +IV)– are present. Gas sensing experiments showed an extraordinary selectivity towards low concentrations of NO2 (ppb level) in comparison to various gas

species, i.e. usually interfering fossil combustion products (SO2, CO, NO, CO2), that were

applied in notably higher concentrations (2 to 50,000 ppm). Among the tested amine species, N-[3-(trimethoxysilyl)-propyl]ethylenediamine (en-APTAS) unified the best performances in both, selectivity and sensitivity towards NO2. It could be shown that fast

response/recovery characteristics, as well as quantitative sensing results can be achieved by illuminating the active sensor surface with visible light in the range of 480-680 nm wavelength.

The simulation of the entire NW-SAM-gas system via DFT methods could identify the crucial parameters for the high selectivity achieved for the SnO2/en-APTAS system towards NO2. It

could be shown that the energy of the lowest unoccupied molecular orbital (LUMO), formed by the en-APTAS/NO2 system, was well below the Fermi Energy of the SAM modified SnO2

NW. In this case, an efficient charge transfer was enabled from the NW via the SAM to NO2.

The consequent depletion of charge carriers (electrons) at the n-type SnO2 surface area

3.1 Selective Sensing 18 resulted in unfavourable energy alignments and thus were consistent with the obtained experimental results, where no response was observed in presence of these species.

Summarized, this paper contains the following aspects on the development of a selective gas sensor system:

 Synthetic realization and chemical analysis of the hybrid sensor material.

 Evaluation of the gas sensing characteristics towards NO2, as well as common

interfering gas species.

 Evaluation of optimal operation conditions in terms of response/recovery characteristics and quantitative sensing results.

 Theoretical identification of crucial parameters to reconstruct the underlying sensing mechanism of the NW-SAM-gas system.

Included paper 2:

Shao, F.; Hoffmann, M. W. G; Prades, J. D; Zamani, R; Arbiol, J.; Morante, J. R.; Varechkina, E.; Rumantseva, M.; Gaskov, A.; Giebelhaus, I.; Fischer, T.; Mathur, S.; Hernandez-Ramirez, F., “Heterostructured p-CuO (Nanoparticle)/n-SnO2 (Nanowire) Devices for Selective H2S

Detection”, Sens. Actuat. B 181, 130-135 (2013). (impact factor: 3.8)

For certain sensor-gas couples selective detection phenomena can be observed even for pure inorganic systems. As a matter of fact, such behaviour is given for p-CuO/n-SnO2

hererostructures when they are exposed to dihydrogen sulphide (H2S) containing

atmospheres. Here, the effect can be attributed to the reaction of p-type CuO with H2S to

form CuS. In absence of H2S, p-CuO forms a p-n heterojunction with n-SnO2 and thus, a

carrier (electrons) depleted zone is formed at their interface. In presence of H2S, p-CuO is

converted to CuS and consequently, the electron depletion zone is vanished at the SnO2

interface. In the study presented in paper 2, a p-CuO NP/SnO2 NW heterostructure was

used with SnO2 NWs as conductive channel. Due to variations of the chemical/electrical

properties of Cu-based semiconductor NPs on the NW surface, the resistance of the SnO2

NWs was gated and a very high response with sharp decrease of resistance in presence of H2S gas could be observed. As the sensitivity of such heterostructures was found to be

clearly above the values observed for bare SnO2 systems, the effect could be attributed of the

p-CuO(CuS)/n-SnO2 heterostructure. Additionally, a good selectivity for H2S could be

observed in comparison to NH3 and CO species. This study describes the influence of a

surface bound component, which undergoes selective interactions with the gas phase (p-CuO), on the monitored semiconductor (n-SnO2). As this principle is also present in case of

19 3 Results and Discussion organic-inorganic hybrid sensor systems, the evaluation of such effects provides an additional understanding for sensing interactions of selective sensor heterostructures. In summary, paper 2 presents the following aspects:

 Experimental realization and analysis of p-CuO NP/n-SnO2 NW heterostructures.

 Evaluation of gas sensing characteristics for single- and multiwire devices.

3.1 Selective Sensing 20

3.1 Selective Sensing 22

3.1 Selective Sensing 48

3.2 Self-Powered Sensing 56

3.2 Self-Powered Sensing

Included Paper 3:

Hoffmann, M. W. G.; Gad, A. E.; Prades, J. D.; Hernandez-Ramirez, F.; Fiz, R.; Shen, H.; Mathur, S., “Solar diode sensor: Sensing mechanism and applications“, Nano Energy 2, 514–522 (2013). (impact factor: 10.2)

Included Paper 4:

Gad, A. E.; Hoffmann, M. W. G.; Hernandez-Ramirez, F.; Prades, J. D.; Shen, H.; Mathur, S., “Coaxial p-Si/n-ZnO nanowire heterostructures for energy and sensing applications”, Mater.

Chem. Phys. 135, 618–622 (2012). (impact factor: 2.1)

The high power consumption of gas sensor devices represents (besides the previously discussed lack of selectivity) a major challenge for the application on mobile platforms. In this section, the concept of a novel multifunctional heterostructure for energy autonomous sensing operations is presented (see paper 3) in which the needed energy for signal generation and surface absorption/desorption activation is harvested solely from solar light. To accomplish this requirement, the CdS@n-ZnO/p-Si hetero structure consists of an energy harvesting (n-ZnO/p-Si diode) and gas sensing unit (CdS@n-ZnO NWs) with a synergetic overlap of the n-ZnO nano structure. Incident light is absorbed at the n-ZnO/p-Si junction and the formed exciton is separated due to the build in potential (Vbi) and creates

an open circuit voltage (Voc) between the p- and n-type terminals. In the sensing unit, CdS, a

semiconductor with relatively narrow band gap (2.4 eV), can absorb incident visible light and transfers the excited electron to the ZnO NW, due to the favourable band alignment. Other than in case of bare wide band gap semiconductors (Eb(ZnO) = 3.4 eV), this synergy

allows for an activation of the gas sensitive material (here ZnO) with solar light. The consequent interaction with the gas atmosphere modulates the intrinsic charge carrier concentration (ND) and hence, modifies Voc. This self-generated signal (Voc) could be utilized

to quantitatively sense oxidative and reductive gases without the need of external energy. The underlying mechanism in terms of gas–material surface interactions and the subsequent changes in the donor density (ND) was proven experimentally.

Summarized, paper 3 presents the following content:

 Design of a multifunctional heterostructure for self-powered gas sensing, unifying energy harvesting und gas sensing unit in a simple and compact layout.

 Evaluation of the gas sensing characteristics for self-powered operation towards oxidizing and reductive gases.

57 3 Results and Discussion  Description of the novel gas sensing mechanism, based on modulations of the

n-ZnO/p-Si build in potential (Vbi).

 Experimental confirmation of the proposed sensing mechanism.

To gain a better understanding of n-ZnO/p-Si based gas sensing systems, a single wire device was realized via focused ion beam (FIB) lithography and is described in paper 4. The p-Si core of the radial heterostructure was accessed by FIB etching and both terminals (n-ZnO shell and p-Si core) were contacted platinum stripes. The single wire configuration is well suited for the evaluation of intrinsic material properties, as statistical dispersion, typical for multi wire or bulk configurations, can be excluded. Additionally, the small dimension of those devices reduces the energy uptake during operation. The single wire device showed diode characteristics and photovoltaic performance under solar light illumination. Modelling of the electrical properties revealed a relatively low resistance in reverse bias, which can be attributed to defects at the radial p-n junction. Although the Voc

signal was not sufficient for gas sensing experiments in a self-powered operation, it was shown that the gas sensing performance could be modulated via the applied external current through the device, attaining a maximal response at low reverse bias currents. The contents of paper 4 can be summarized as following:

 Realization of a single wire n-ZnO/p-Si device via FIB-assisted lithography.  Evaluation and modelling of the electrical and photovoltaic device properties.  Optimization of the gas sensor working conditions.

3.2 Self-Powered Sensing 58

3.2 Self-Powered Sensing 76

3.3 Combination of Selective and Self-Powered Sensing 82

3.3 Combination of Selective and Self-Powered Sensing

Included Paper 5:

Hoffmann, M. W. G.; Mayrhofer, L.; Casals, O.; Caccomo, L.; Hernandez-Ramirez, F.; Moseler, M.; Waag, A.; Shen, H.; Prades, J. D., “Highly selective and self-powered gas sensor enabled via organic surface functionalization”, Adv. Mater. (2014). doi: 10.1002/adma.201403073 (impact factor: 15.4)

To fulfil the above discussed requirements for mobile sensing devices on reliable gas species identification and low power consumption, the concepts described in sections 1 and 2 were combined to realize a highly selective and self-powered sensor. Paper 5 describes the resulting device starting from aspects of processing and study of the gas sensing characteristics to the evaluation of the sensing mechanism. Following the concept introduced in section 2, the sensor unit is based on p-Si/n-ZnO diodes to generate the sensor signal (Voc) under illumination conditions. In order to amplify the Voc signal even under low

illumination intensity, the device was designed with several (9, 16, 26) diodes connected in series. To realize this configuration, photolithographic methods and a novel procedure for site selective growth of n-ZnO NWs were applied. The resulting devices generated a very high Voc signal of up to 1.84 V (26 diodes in series; Voc,exp = 72±6 mV/diode) upon

illumination with simulated sunlight.

Simulations of the device band structure and optoelectronic response upon illumination could describe the characteristics of the p-Si/n-ZnO solar cell. The p-Si bulk was identified to be the exclusive light absorbing material. A shunt resistance of about 30 Ωcm2 _{at the}

p-Si/n-ZnO interface was found to reduce the experimentally observed Voc in comparison to the

theoretical value (Voc.ideal = 0.479 mV/diode).

Organic methoxysilanes were used to bind to the surface of n-ZnO NWs and to form a functional SAM surface for sensor-gas interactions, as it was discussed in section 1. In order to study the influence of functional organic groups of the SAMs on the gas sensing characteristics, amine and thiole terminated methoxysilanes were used. The successful immobilization on the n-ZnO surface was proven by spatially resolved Auger spectroscopy indexing the characteristic atomic components. Sensing experiments showed a high selectivity towards NO2 in comparison to other typical interfering gases (SO2, CO, NH3).

Notably, a contradictory signal directions for amine and thiole functionalized sensors were observed, proving the crucial influence of the organic functional groups on the senor characteristics. Both sensors were capable of detecting low NO2 concentrations down to

250 ppb. In case of amine terminated devices the linear increase of response with increasing NO2 concentration further allowed for quantitative gas sensing.

83 3 Results and Discussion DFT simulations of the NO2–SAM geometries for thiole and amine terminated sensors

revealed diverging binding characteristics with stronger binding energy (EB) in the amine

case (EB = -0.44 eV) compared to thiole functionalized NWs (EB = -0.22 eV). Furthermore, the

observed diverging energetic changes of the SAM frontier orbitals upon NO2 binding could

give a possible explanation for the experimentally obtained contrary responses of amine and thiole functionalized sensors.

Summarized, this paper contains the following aspects on the development of a selective self-powered sensor:

 Fabrication of a sensor system consisting of multiple diodes in series by using lithographic methods and site selective growth of aligned n-ZnO NWs.

 Evaluation of the electrical device characteristics via impedance spectroscopy  Realization and evaluation of amine and thiole terminated SAMs on n-ZnO NWs  Evaluation of the gas sensor characteristics in terms of selectivity and sensitivity  Theoretical evaluation of the sensor-gas binding interactions and consequent

variations of the electrical structure within the SAMs in terms of their ionization potential and electron affinity.

3.3 Combination of Selective and Self-Powered Sensing 84

114

4

Conclusion

Within the here presented thesis, a device novel concept for selective sensing of NO2 without

the need of external energy input was developed. The realization of the thesis outline required an interdisciplinary approach involving fields of chemistry, physics and electrical engineering. It was demonstrated that the specific unification of functional inorganic (hetero) materials with defined organic functionalities can lead to powerful hybrid devices with performances far beyond their singular units. It was described, that the understanding of the observed sensing characterizations of these hybrid devices cannot be described by a consideration of isolated inorganic and organic stakes. Moreover, an understanding of the complex aggregated system from SAM-gas binding and electronic modulations within the inorganic as well as organic structures are needed to understand the sensor-gas interaction mechanisms that can lead to selective gas sensing properties.

Fig. 4.1: (left) Summary of sensing responses observed for the SAM-NW hybrid sensor and (right)

DFT simulations of the charge transfer dynamics upon NO2-SAM binding.

Section 1 described the concept of a selective sensor based on SAM modified inorganic NWs. An extraordinary sensing performance in terms of selectivity and sensitivity towards NO2

could be achieved when compared to known semiconductor based systems. For the first time, the underlying mechanism of sensor-gas binding interactions and consequent charge transfer sequences between the organic and inorganic components were analysed via DFT methods and revealed a consistence of experimental and theoretical evaluations. The

-20 0 20 40 2,000 2,100 2,200 5 0 ,0 0 0 p p m 2 0 0 p p m 1 0 0 p p m 2 0 0 p p m 2 p p m 4 p p m NO NO₂ SO₂ EtOH NH₃ CO CO₂ S / % 0.4 p p m NO2 SO2 NO